[PATCH] dmaengine: dma_slave_caps: remove sg entries

[PATCH] dmaengine: dma_slave_caps: remove sg entries

[Vinod Koul]

As pointed by Russell in [1], the sg properties are already availble in struct device,
so no need to duplicate here.

[1]: http://marc.info/?l=linux-omap&m=137416733628831

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
---
 include/linux/dmaengine.h |    8 --------
 1 files changed, 0 insertions(+), 8 deletions(-)

diff --git a/include/linux/dmaengine.h b/include/linux/dmaengine.h
index 5692bc3..4310b89 100644
--- a/include/linux/dmaengine.h
+++ b/include/linux/dmaengine.h
@@ -380,11 +380,6 @@ struct dma_slave_config {
  * 	should be checked by controller as well
  * @cmd_pause: true, if pause and thereby resume is supported
  * @cmd_terminate: true, if terminate cmd is supported
- *
- * @max_sg_nr: maximum number of SG segments supported
- * 	0 for no maximum
- * @max_sg_len: maximum length of a SG segment supported
- * 	0 for no maximum
  */
 struct dma_slave_caps {
 	u32 src_addr_widths;
@@ -392,9 +387,6 @@ struct dma_slave_caps {
 	u32 directions;
 	bool cmd_pause;
 	bool cmd_terminate;
-
-	u32 max_sg_nr;
-	u32 max_sg_len;
 };
 
 static inline const char *dma_chan_name(struct dma_chan *chan)
-- 
1.7.0.4

Re: [PATCH] dmaengine: dma_slave_caps: remove sg entries

[Lars-Peter Clausen]

On 09/02/2013 02:21 PM, Vinod Koul wrote:
> As pointed by Russell in [1], the sg properties are already availble in struct device,
> so no need to duplicate here.
> 
> [1]: http://marc.info/?l=linux-omap&m=137416733628831

Hm, I guess that will work as well, but shouldn't we update the drivers that
already implementing this interface first?

- Lars

Re: [PATCH] dmaengine: dma_slave_caps: remove sg entries

[Vinod Koul]

On Mon, Sep 02, 2013 at 04:52:06PM +0200, Lars-Peter Clausen wrote:
> On 09/02/2013 02:21 PM, Vinod Koul wrote:
> > As pointed by Russell in [1], the sg properties are already availble in struct device,
> > so no need to duplicate here.
> > 
> > [1]: http://marc.info/?l=linux-omap&m=137416733628831
> 
> Hm, I guess that will work as well, but shouldn't we update the drivers that
> already implementing this interface first?
Well thats why I cced you :). I have fixed up the pl330 driver. Have you used
these values in any of the ASoC drivers?

~Vinod
-- 

Re: [PATCH] dmaengine: dma_slave_caps: remove sg entries

[Vinod Koul]

On Mon, Sep 02, 2013 at 07:10:52PM +0200, Lars-Peter Clausen wrote:
> On 09/02/2013 06:12 PM, Vinod Koul wrote:
> > On Mon, Sep 02, 2013 at 04:52:06PM +0200, Lars-Peter Clausen wrote:
> >> On 09/02/2013 02:21 PM, Vinod Koul wrote:
> >>> As pointed by Russell in [1], the sg properties are already availble in struct device,
> >>> so no need to duplicate here.
> >>>
> >>> [1]: http://marc.info/?l=linux-omap&m=137416733628831
> >>
> >> Hm, I guess that will work as well, but shouldn't we update the drivers that
> >> already implementing this interface first?
> > Well thats why I cced you :). I have fixed up the pl330 driver. Have you used
> > these values in any of the ASoC drivers?
> 
> Not in upstream yet, if the pl330 is fixed we are good.
Thanks for the confirmation. I had a week old Takashi's tree, didnt find
anything there, but still wanted your ack.

Now on pl330 here is the update

---

From: Vinod Koul <vinod.koul@intel.com>
Date: Mon, 2 Sep 2013 21:54:48 +0530
Subject: [PATCH] dmaengine: pl330: use dma_set_max_seg_size to set the sg limit

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
---
compile tested only

 drivers/dma/pl330.c |   15 ++++++++-------
 1 files changed, 8 insertions(+), 7 deletions(-)

diff --git a/drivers/dma/pl330.c b/drivers/dma/pl330.c
index 36ed301..a562d24 100644
--- a/drivers/dma/pl330.c
+++ b/drivers/dma/pl330.c
@@ -2886,13 +2886,6 @@ static int pl330_dma_device_slave_caps(struct dma_chan *dchan,
 	caps->cmd_pause = false;
 	caps->cmd_terminate = true;
 
-	/*
-	 * This is the limit for transfers with a buswidth of 1, larger
-	 * buswidths will have larger limits.
-	 */
-	caps->max_sg_len = 1900800;
-	caps->max_sg_nr = 0;
-
 	return 0;
 }
 
@@ -3017,6 +3010,14 @@ pl330_probe(struct amba_device *adev, const struct amba_id *id)
 			"unable to register DMA to the generic DT DMA helpers\n");
 		}
 	}
+	/*
+	 * This is the limit for transfers with a buswidth of 1, larger
+	 * buswidths will have larger limits.
+	 */
+	ret = dma_set_max_seg_size(&adev->dev, 1900800);
+	if (ret)
+		dev_err(&adev->dev, "unable to set the seg size\n");
+
 
 	dev_info(&adev->dev,
 		"Loaded driver for PL330 DMAC-%d\n", adev->periphid);
-- 
1.7.0.4

Re: [PATCH] dmaengine: dma_slave_caps: remove sg entries

[Lars-Peter Clausen]

On 09/02/2013 06:12 PM, Vinod Koul wrote:
> On Mon, Sep 02, 2013 at 04:52:06PM +0200, Lars-Peter Clausen wrote:
>> On 09/02/2013 02:21 PM, Vinod Koul wrote:
>>> As pointed by Russell in [1], the sg properties are already availble in struct device,
>>> so no need to duplicate here.
>>>
>>> [1]: http://marc.info/?l=linux-omap&m=137416733628831
>>
>> Hm, I guess that will work as well, but shouldn't we update the drivers that
>> already implementing this interface first?
> Well thats why I cced you :). I have fixed up the pl330 driver. Have you used
> these values in any of the ASoC drivers?

Not in upstream yet, if the pl330 is fixed we are good.

- Lars